1. Field of the Invention
The present invention relates to disk drives. In particular, the present invention relates to a disk drive that modifies estimated seek times for a rotational position optimization (RPO) algorithm based on change in estimated seek time parameter.
2. Description of the Prior Art
A disk drive may employ an RPO algorithm in order to execute commands in an order which minimizes the seek latency of the head as well as the rotational latency of the disk. After executing a current command, the RPO algorithm will typically evaluate various parameters to select the next command that minimizes the access time with respect to the radial and circumferential location of the head. The seek latency of the head (the time required to move the head from a current track to a new track) has typically been determined by evaluating a small number of disk drives to establish nominal seek profiles for a family of disk drives. Each individual disk drive is then manufactured with the nominal seek profiles which may lead to sub-optimal performance since the nominal seek profiles are selected to account for worst case conditions. U.S. patent application Ser. No. 10/060,881 Pub. No. US 2002/0131195 discloses a method for calibrating the seek profiles for each individual disk drive in a family of disk drives during manufacturing, as well as updating the seek profiles “in the field” to account for changes in the disk drive that occur over time.
The method disclosed in the '881 patent application includes a manufacturing process for each individual disk drive wherein the seek time of the head to travel a distance D is measured over multiple seeks and statistically averaged to establish an initial seek profile. A problem with this technique, however, is it increases the manufacturing time significantly due to the multiple seeks performed for each seek distance D, as well as the numerous seek distances that must be calibrated. While in the field, the '881 patent application updates the seek profiles for each individual disk drive by statistically averaging the actual seek times with the current seek profiles. A problem with this technique, however, is the statistical averaging algorithm must have a very slow response in order to filter noise. This means the seek profiles will be updated slowly in response to changes in the disk drive leading to sub-optimal tracking of faster deviations.
There is, therefore, a need to customize the RPO algorithm for each individual disk drive without significantly increasing the manufacturing time. There is also a need to modify the RPO algorithm to quickly track changes in the operating characteristics of each individual disk drive while in the field.